The need for reliable and valid measures of quality of care increases in importance as new attempts to control the rate of increase of American health care costs are tested and implemented. Our research team has developed a nationally representative data base for reliably and validly measuring quality of care using detailed explicit and implicit review of the medical record and secondary data as the sources of data. The study sample consists of 14,012 Medicare patients 65 years or older who are hospitalized with one of five diseases: congestive heart failure, myocardial infarction, pneumonia, stroke, or hip fracture. The nationally representative sample comes from 297 hospitals in five states. We have under separate funding already answered a series of important policy questions about the effects of the DRG-based prospective payment system on quality of care. We now have the opportunity to use this unique and already established data base to evaluate a series of specific and testable methodologic and policy questions relating to quality of care for hospitalized Medicare patients. Specifically, we propose to answer a series of questions about clinical epidemiology, quality of care methods, and health policies. We are proposing to use the already existing data base to address 9 problems in the measurement of quality of care for hospitalized patients. The topics pertain to the identification of unique clinical characteristics of each disease; the analysis of quality for patients with quick hospital deaths; problems associated with missing data; the incidence and treatment of adverse outcomes; the use of the chest x-ray report as a measure of patient sickness; the predictors of death after hospitalization; the predictors of readmissions after hospitalization; the relationships between implicit and explicit measures of quality; and the use of existing data that distinguishes better and worse hospitals to understand how hospital-based care could improve and by how much. The products of these projects will be a series of manuscripts documenting the analyses and the implications of these analyses for future measurement of quality of care.